get-claudia 1.53.0 → 1.53.2

package/CHANGELOG.md

@@ -2,6 +2,32 @@
 
 All notable changes to Claudia will be documented in this file.
 
+## 1.53.2 (2026-03-04)
+
+### Re-enable Gmail and Calendar MCPs
+
+Gmail and Calendar MCP servers were disabled in v1.53.1 as a workaround for Claude Code bug #17962 (multiple stdio servers). Multiple stdio servers now work reliably, so these are re-enabled.
+
+- **Gmail and Calendar enabled by default** -- `.mcp.json.example` ships `gmail` and `google-calendar` as active entries (no `_disabled_` prefix). Both require Google Cloud credentials (see Google Integration Setup in CLAUDE.md).
+- **Restore function re-enables Gmail/Calendar** -- `restoreMcpServers()` now also restores `gmail` and `google-calendar` from `_disabled_mcpServers` on upgrade. Additionally handles `_disabled_`-prefixed keys in `mcpServers` itself.
+- **Diagnose skill updated** -- Removed multiple-stdio warning (Step 1b) and the advice to disable Gmail/Calendar. Step 1b now simply lists active servers without warning about conflicts.
+- **Removed stdio_warning from config notes** -- `.mcp.json.example` no longer warns about multiple stdio servers or suggests disabling Gmail/Calendar.
+
+## 1.53.1 (2026-03-04)
+
+### The One-Stdio Fix: Memory Tools Actually Connect
+
+The daemon was healthy, all 26 tools registered, preflight passed, but tools silently vanished from Claude Code's palette. Root cause: `.mcp.json.example` shipped with 4 active stdio MCP servers. Claude Code bug [#17962](https://github.com/anthropics/claude-code/issues/17962) means only one stdio server connects reliably. Additionally, the daemon's JSON schemas used union type shorthand (`["integer","string"]`) that Claude Code's Zod validator may reject ([#10031](https://github.com/anthropics/claude-code/issues/10031)).
+
+- **Disable extra stdio servers by default** -- `.mcp.json.example` now ships gmail and google-calendar as `_disabled_` prefixed keys. Only `claudia-memory` is active as stdio. Matches the v1.49.0 pattern that worked.
+- **Rube switched to HTTP transport** -- Rube (Composio) entry changed from `command: npx` (stdio) to `type: http` with `url: https://mcp.composio.dev`. No stdio conflict, no slot consumed.
+- **Fix JSON Schema union types** -- Tool schemas changed from `["integer","string"]` and `["array","string"]` to simple `"string"` and `"array"`. The `_coerce_int()` and `_coerce_arg()` helpers already handle type conversion at runtime, so the unions were unnecessary and potentially caused all tools to vanish.
+- **Multi-stdio detection in installer** -- New `warnMultipleStdioServers()` warns during install if >1 stdio server is active. New "MCP Config" progress step shows stdio server count.
+- **Post-install component checklist** -- Installer now shows a component summary (Personality & Skills, Memory Daemon, MCP Config, Stdio Servers) with status indicators after install/upgrade.
+- **Diagnose skill: Step 1b** -- New diagnostic step counts active stdio servers and flags multiples. New common issue entry for "Multiple stdio MCP servers" with fix instructions.
+- **Fix `restoreMcpServers()`** -- Restore function no longer re-enables gmail/google-calendar as active stdio servers on upgrade. Only `claudia-memory` is restored from `_disabled_mcpServers`.
+- **Closed GitHub issues** -- [#7](https://github.com/kbanc85/claudia/issues/7) (migration 20 UNIQUE constraint) and [#8](https://github.com/kbanc85/claudia/issues/8) (schema.sql missing from npm) were both fixed in earlier releases by the removal of the Node.js CLI.
+
 ## 1.53.0 (2026-03-04)
 
 ### The Reliability Release: Self-Diagnosing MCP Daemon

package/bin/index.js

@@ -96,6 +96,7 @@ const STEPS = [
   { id: 'models',      label: 'AI Models' },
   { id: 'memory',      label: 'Memory System' },
   { id: 'daemon',      label: 'Memory Daemon' },
+  { id: 'mcp',         label: 'MCP Config' },
   { id: 'vault',       label: 'Obsidian Vault' },
   { id: 'health',      label: 'Health Check' },
 ];
@@ -818,9 +819,23 @@ async function main() {
       await ensureLaunchAgent(venvPython, targetPath);
     }
 
-    // Step 5 (vault): handled below
+    // MCP Config step: verify .mcp.json is correct and check stdio server count
+    renderer.update('mcp', 'active', 'checking .mcp.json...');
+    const mcpCheckResult = checkMcpConfig(targetPath);
+    if (mcpCheckResult.hasDaemon && mcpCheckResult.stdioCount <= 1) {
+      renderer.update('mcp', 'done', `claudia-memory configured${mcpCheckResult.stdioCount === 1 ? '' : ' (no stdio servers?)'}`);
+      if (!supportsInPlace) renderer.appendLine('mcp', 'done', 'claudia-memory configured');
+    } else if (mcpCheckResult.hasDaemon && mcpCheckResult.stdioCount > 1) {
+      renderer.update('mcp', 'warn', `${mcpCheckResult.stdioCount} stdio servers (only 1 reliable)`);
+      if (!supportsInPlace) renderer.appendLine('mcp', 'warn', `${mcpCheckResult.stdioCount} stdio servers`);
+    } else {
+      renderer.update('mcp', 'warn', 'claudia-memory not in .mcp.json');
+      if (!supportsInPlace) renderer.appendLine('mcp', 'warn', 'daemon not configured');
+    }
+
+    // Vault step: handled below
 
-    // Step 6: Health Check -- check daemon health endpoint or verify daemon can import
+    // Health Check: check daemon health endpoint or verify daemon can import
     renderer.update('health', 'active', 'verifying...');
     let healthOk = false;
 
@@ -950,16 +965,34 @@ async function main() {
 
     console.log('');
     console.log(`${colors.dim}${'━'.repeat(46)}${colors.reset}`);
-    console.log(` ${colors.bold}Done!${colors.reset} Open Claude Code:`);
+
+    // Post-install component summary
+    const mcpCheck = checkMcpConfig(targetPath);
+    const components = [
+      { name: 'Personality & Skills', ok: true },
+      { name: 'Memory Daemon', ok: memoryInstalled },
+      { name: 'MCP Config', ok: mcpCheck.hasDaemon },
+    ];
+
+    console.log(` ${colors.bold}Components:${colors.reset}`);
+    for (const c of components) {
+      if (c.ok && !c.warn) {
+        console.log(`   ${colors.green}✓${colors.reset} ${c.name}`);
+      } else if (c.warn) {
+        console.log(`   ${colors.yellow}⚠${colors.reset} ${c.name} ${colors.dim}${c.warn}${colors.reset}`);
+      } else {
+        console.log(`   ${colors.yellow}○${colors.reset} ${c.name} ${colors.dim}(not ready)${colors.reset}`);
+      }
+    }
+
+    console.log('');
+    console.log(` ${colors.bold}Next:${colors.reset} Open Claude Code:`);
     console.log(`   ${colors.cyan}${cdCmd}claude${colors.reset}`);
 
     if (!memoryInstalled) {
       console.log('');
       console.log(` ${colors.dim}Memory requires Python 3.10+, the claudia-memory daemon, and Ollama.${colors.reset}`);
       console.log(` ${colors.dim}Re-run setup after installing Python and Ollama.${colors.reset}`);
-    } else {
-      console.log('');
-      console.log(` ${colors.dim}Memory system ready. Claudia will remember across sessions.${colors.reset}`);
     }
 
     console.log('');
@@ -969,9 +1002,10 @@ async function main() {
 
 /**
  * Restore MCP servers that were moved to _disabled_mcpServers by earlier versions.
- * - Gmail/Calendar: v1.51.9-v1.51.12 treated them as legacy, now restored.
  * - claudia-memory: v1.51.13+ treated the daemon as legacy (replaced by CLI),
  *   but MCP is the primary memory interface as of v1.51.22.
+ * - Gmail/Calendar: v1.53.1 disabled these due to Claude Code bug #17962,
+ *   but multiple stdio servers now work reliably. Restore them.
  */
 function restoreMcpServers(targetPath) {
   const mcpPath = join(targetPath, '.mcp.json');
@@ -983,7 +1017,8 @@ function restoreMcpServers(targetPath) {
     if (!config._disabled_mcpServers) return;
     if (!config.mcpServers) config.mcpServers = {};
 
-    const toRestore = ['gmail', 'google-calendar', 'claudia-memory', 'claudia_memory'];
+    // Restore all previously disabled servers (memory, gmail, google-calendar)
+    const toRestore = ['claudia-memory', 'claudia_memory', 'gmail', 'google-calendar'];
     let changed = false;
     const restored = [];
 
@@ -991,6 +1026,7 @@ function restoreMcpServers(targetPath) {
       if (config._disabled_mcpServers[key] && !config.mcpServers[key]) {
         const serverConfig = { ...config._disabled_mcpServers[key] };
         delete serverConfig._replaced_by;
+        delete serverConfig._warning;
         config.mcpServers[key] = serverConfig;
         delete config._disabled_mcpServers[key];
         changed = true;
@@ -998,6 +1034,21 @@ function restoreMcpServers(targetPath) {
       }
     }
 
+    // Also rename _disabled_ prefixed keys in mcpServers itself
+    for (const key of Object.keys(config.mcpServers)) {
+      if (key.startsWith('_disabled_')) {
+        const realKey = key.replace('_disabled_', '');
+        if (['gmail', 'google-calendar'].includes(realKey) && !config.mcpServers[realKey]) {
+          const serverConfig = { ...config.mcpServers[key] };
+          delete serverConfig._warning;
+          config.mcpServers[realKey] = serverConfig;
+          delete config.mcpServers[key];
+          changed = true;
+          restored.push(realKey);
+        }
+      }
+    }
+
     // Clean up _disabled_mcpServers if it's now empty
     if (config._disabled_mcpServers && Object.keys(config._disabled_mcpServers).length === 0) {
       delete config._disabled_mcpServers;
@@ -1055,6 +1106,28 @@ function ensureDaemonMcpConfig(targetPath, venvPythonPath) {
   writeFileSync(mcpPath, JSON.stringify(config, null, 2) + '\n');
 }
 
+/**
+ * Check .mcp.json configuration and return status.
+ * Returns { hasDaemon, stdioCount, stdioServers }.
+ */
+function checkMcpConfig(targetPath) {
+  const mcpPath = join(targetPath, '.mcp.json');
+  if (!existsSync(mcpPath)) return { hasDaemon: false, stdioCount: 0, stdioServers: [] };
+  try {
+    const config = JSON.parse(readFileSync(mcpPath, 'utf-8'));
+    const servers = config.mcpServers || {};
+    const hasDaemon = !!servers['claudia-memory'];
+    const stdioServers = Object.entries(servers)
+      .filter(([key]) => !key.startsWith('_'))
+      .filter(([, val]) => !val._disabled && (!val.type || val.type === 'stdio'))
+      .map(([key]) => key);
+    return { hasDaemon, stdioCount: stdioServers.length, stdioServers };
+  } catch {
+    return { hasDaemon: false, stdioCount: 0, stdioServers: [] };
+  }
+}
+
+
 /**
  * Register (or update) the macOS LaunchAgent for the standalone daemon.
  * The standalone daemon runs 24/7 for scheduled jobs (consolidation, decay, vault sync).

package/memory-daemon/claudia_memory/mcp/server.py

@@ -1614,7 +1614,7 @@ async def list_tools() -> ListToolsResult:
                         "default": "fact",
                     },
                     "about": {
-                        "type": ["array", "string"],
+                        "type": "array",
                         "items": {"type": "string"},
                         "description": "Entity names this memory relates to (people, projects, etc.)",
                     },
@@ -1668,12 +1668,12 @@ async def list_tools() -> ListToolsResult:
                         "description": "What to search for (required unless ids is provided)",
                     },
                     "limit": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "description": "Maximum number of results",
                         "default": 10,
                     },
                     "types": {
-                        "type": ["array", "string"],
+                        "type": "array",
                         "items": {"type": "string"},
                         "description": "Filter by memory types (fact, preference, observation, learning, commitment)",
                     },
@@ -1687,7 +1687,7 @@ async def list_tools() -> ListToolsResult:
                         "default": False,
                     },
                     "ids": {
-                        "type": ["array", "string"],
+                        "type": "array",
                         "items": {"type": "integer"},
                         "description": "Fetch specific memories by ID (skips search). Use after a compact search to get full content.",
                     },
@@ -1711,7 +1711,7 @@ async def list_tools() -> ListToolsResult:
                         "description": "Name of the person, project, or entity",
                     },
                     "limit": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "description": "Maximum number of memories to return",
                         "default": 20,
                     },
@@ -1798,7 +1798,7 @@ async def list_tools() -> ListToolsResult:
                 "type": "object",
                 "properties": {
                     "episode_id": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "description": "Episode ID from buffer_turn calls during the session",
                     },
                     "narrative": {
@@ -1812,7 +1812,7 @@ async def list_tools() -> ListToolsResult:
                         ),
                     },
                     "facts": {
-                        "type": ["array", "string"],
+                        "type": "array",
                         "items": {
                             "type": "object",
                             "properties": {
@@ -1846,7 +1846,7 @@ async def list_tools() -> ListToolsResult:
                         "description": "Structured facts, preferences, observations, learnings extracted from the session",
                     },
                     "commitments": {
-                        "type": ["array", "string"],
+                        "type": "array",
                         "items": {
                             "type": "object",
                             "properties": {
@@ -1874,7 +1874,7 @@ async def list_tools() -> ListToolsResult:
                         "description": "Commitments or promises made during the session",
                     },
                     "entities": {
-                        "type": ["array", "string"],
+                        "type": "array",
                         "items": {
                             "type": "object",
                             "properties": {
@@ -1895,7 +1895,7 @@ async def list_tools() -> ListToolsResult:
                         "description": "New or updated entities mentioned during the session",
                     },
                     "relationships": {
-                        "type": ["array", "string"],
+                        "type": "array",
                         "items": {
                             "type": "object",
                             "properties": {
@@ -1909,12 +1909,12 @@ async def list_tools() -> ListToolsResult:
                         "description": "Relationships between entities observed during the session",
                     },
                     "key_topics": {
-                        "type": ["array", "string"],
+                        "type": "array",
                         "items": {"type": "string"},
                         "description": "Main topics discussed in the session",
                     },
                     "reflections": {
-                        "type": ["array", "string"],
+                        "type": "array",
                         "items": {
                             "type": "object",
                             "properties": {
@@ -1971,7 +1971,7 @@ async def list_tools() -> ListToolsResult:
                         "description": "Semantic search query (optional). If omitted, returns recent high-importance reflections.",
                     },
                     "types": {
-                        "type": ["array", "string"],
+                        "type": "array",
                         "items": {
                             "type": "string",
                             "enum": ["observation", "pattern", "learning", "question"],
@@ -1983,7 +1983,7 @@ async def list_tools() -> ListToolsResult:
                         "description": "Filter to reflections about a specific entity (optional)",
                     },
                     "limit": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "default": 10,
                         "description": "Maximum results to return",
                     },
@@ -1994,7 +1994,7 @@ async def list_tools() -> ListToolsResult:
                         "description": "Action: get (list), search (semantic), update, delete",
                     },
                     "reflection_id": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "description": "Reflection ID (required for update/delete actions)",
                     },
                     "content": {
@@ -2019,7 +2019,7 @@ async def list_tools() -> ListToolsResult:
                 "type": "object",
                 "properties": {
                     "operations": {
-                        "type": ["array", "string"],
+                        "type": "array",
                         "description": "Array of operations to execute in order",
                         "items": {
                             "type": "object",
@@ -2181,7 +2181,7 @@ async def list_tools() -> ListToolsResult:
                         "description": "List of entity names to summarize",
                     },
                     "top_facts_limit": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "description": "Maximum top facts per entity (default 5)",
                         "default": 5,
                     },
@@ -2234,7 +2234,7 @@ async def list_tools() -> ListToolsResult:
                         "description": "Name of the person to project health for",
                     },
                     "days_ahead": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "description": "How far ahead to project (default 30)",
                         "default": 30,
                     },
@@ -2261,7 +2261,7 @@ async def list_tools() -> ListToolsResult:
                         "description": "Which temporal query to run",
                     },
                     "days": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "description": "Look ahead window in days (for upcoming, default 14)",
                         "default": 14,
                     },
@@ -2279,7 +2279,7 @@ async def list_tools() -> ListToolsResult:
                         "description": "Entity name (for timeline, or optional filter for since)",
                     },
                     "limit": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "description": "Maximum results",
                         "default": 50,
                     },
@@ -2319,12 +2319,12 @@ async def list_tools() -> ListToolsResult:
                         "description": "Second entity (for path operation)",
                     },
                     "max_depth": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "description": "Maximum hops to search (for path, default 4)",
                         "default": 4,
                     },
                     "min_connections": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "description": "Minimum relationships to be a hub (for hubs, default 5)",
                         "default": 5,
                     },
@@ -2334,7 +2334,7 @@ async def list_tools() -> ListToolsResult:
                         "description": "Filter by entity type (for hubs)",
                     },
                     "days": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "description": "Days without activity (for dormant, default 60)",
                         "default": 60,
                     },
@@ -2344,7 +2344,7 @@ async def list_tools() -> ListToolsResult:
                         "default": 0.3,
                     },
                     "limit": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "description": "Maximum results to return",
                         "default": 20,
                     },
@@ -2384,7 +2384,7 @@ async def list_tools() -> ListToolsResult:
                         "description": "Entity description (for create)",
                     },
                     "aliases": {
-                        "type": ["array", "string"],
+                        "type": "array",
                         "items": {"type": "string"},
                         "description": "Alternative names (for create)",
                     },
@@ -2393,20 +2393,20 @@ async def list_tools() -> ListToolsResult:
                         "description": "Search query (for search)",
                     },
                     "types": {
-                        "type": ["array", "string"],
+                        "type": "array",
                         "items": {"type": "string"},
                         "description": "Filter by entity types (for search)",
                     },
                     "source_id": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "description": "Entity ID to merge FROM (for merge)",
                     },
                     "target_id": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "description": "Entity ID to merge INTO (for merge)",
                     },
                     "entity_id": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "description": "Entity ID (for delete)",
                     },
                     "reason": {
@@ -2414,7 +2414,7 @@ async def list_tools() -> ListToolsResult:
                         "description": "Reason for merge or delete",
                     },
                     "entities": {
-                        "type": ["array", "string"],
+                        "type": "array",
                         "items": {"type": "string"},
                         "description": "Entity names for overview",
                     },
@@ -2429,7 +2429,7 @@ async def list_tools() -> ListToolsResult:
                         "default": True,
                     },
                     "limit": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "description": "Maximum results (for search, default 10)",
                         "default": 10,
                     },
@@ -2492,7 +2492,7 @@ async def list_tools() -> ListToolsResult:
                         "description": "Which modification to perform",
                     },
                     "memory_id": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "description": "Memory ID (for correct, invalidate)",
                     },
                     "correction": {
@@ -2545,7 +2545,7 @@ async def list_tools() -> ListToolsResult:
                         "description": "What the assistant said (for buffer)",
                     },
                     "episode_id": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "description": "Episode ID from previous buffer call (for buffer)",
                     },
                     "source": {
@@ -2598,12 +2598,12 @@ async def list_tools() -> ListToolsResult:
                         "description": "Brief summary (for store)",
                     },
                     "about": {
-                        "type": ["array", "string"],
+                        "type": "array",
                         "items": {"type": "string"},
                         "description": "Entity names this document relates to",
                     },
                     "memory_ids": {
-                        "type": ["array", "string"],
+                        "type": "array",
                         "items": {"type": "integer"},
                         "description": "Memory IDs to link (for store)",
                     },
@@ -2620,7 +2620,7 @@ async def list_tools() -> ListToolsResult:
                         "description": "Filter by entity (for search)",
                     },
                     "limit": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "description": "Maximum results (for search, default 20)",
                         "default": 20,
                     },
@@ -2647,15 +2647,15 @@ async def list_tools() -> ListToolsResult:
                         "description": "Which provenance operation to perform",
                     },
                     "memory_id": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "description": "Memory ID (for trace, or audit by memory)",
                     },
                     "entity_id": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "description": "Entity ID (for audit by entity)",
                     },
                     "limit": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "description": "Maximum audit entries (default 20)",
                         "default": 20,
                     },
@@ -2712,7 +2712,7 @@ async def list_tools() -> ListToolsResult:
                         "description": "What context to build for",
                     },
                     "token_budget": {
-                        "type": ["integer", "string"],
+                        "type": "string",
                         "description": "Maximum tokens (default: 8000)",
                     },
                     "include_sacred": {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "get-claudia",
-  "version": "1.53.0",
+  "version": "1.53.2",
   "description": "An AI assistant who learns how you work.",
   "keywords": [
     "claudia",

package/template-v2/.claude/skills/diagnose/SKILL.md

@@ -36,6 +36,24 @@ If the Python binary in the `command` field doesn't exist:
 ls -la ~/.claudia/daemon/venv/bin/python 2>/dev/null || echo "Daemon venv not found"
 ```
 
+### Step 1b: Check Active MCP Servers
+
+List all active MCP servers (entries without `_disabled` prefix):
+
+```bash
+python3 -c "
+import json
+c = json.load(open('.mcp.json'))
+servers = c.get('mcpServers', {})
+active = [k for k, v in servers.items() if not k.startswith('_')]
+stdio = [k for k in active if servers[k].get('type', 'stdio') == 'stdio']
+http = [k for k in active if servers[k].get('type') == 'http']
+print(f'Active servers ({len(active)}): {chr(44).join(active)}')
+print(f'  stdio: {chr(44).join(stdio) or \"none\"}')
+print(f'  http: {chr(44).join(http) or \"none\"}')
+"
+```
+
 ### Step 2: Run Preflight Check
 
 The daemon has a built-in preflight validator that tests all 11 startup steps:
@@ -108,6 +126,7 @@ Format the diagnosis as:
 | Component | Status | Details |
 |-----------|--------|---------|
 | .mcp.json config | ✅/❌ | [daemon entry present/missing] |
+| Active MCP servers | ✅/⚠️ | [list of active servers] |
 | Daemon Python binary | ✅/❌ | [path exists/missing] |
 | Preflight | ✅/❌ | [all passed / N failures] |
 | Session manifest | ✅/❌ | [running/died/never started] |

package/template-v2/.mcp.json.example

@@ -17,25 +17,26 @@
       "_setup": "Requires your own Google Cloud credentials. See Google Integration Setup in CLAUDE.md, then run: npx @gongrzhe/server-calendar-autoauth-mcp auth"
     },
     "rube": {
-      "command": "npx",
-      "args": ["-y", "@composio/rube-mcp"],
-      "env": {
-        "COMPOSIO_API_KEY": ""
+      "type": "http",
+      "url": "https://mcp.composio.dev",
+      "headers": {
+        "x-composio-api-key": ""
       },
-      "_setup": "Connect 500+ apps via Rube: Slack, Notion, Drive, GitHub, Granola, Otter.ai, Jira, Linear, Airtable, HubSpot, Stripe, and more. 1) Sign up free at https://rube.app 2) Connect your apps in Rube's marketplace 3) Copy your API key from the dashboard 4) Paste it in COMPOSIO_API_KEY above 5) Restart Claude Code. See Rube section in CLAUDE.md for full guide."
+      "_description": "Rube by Composio -- 500+ app integrations (HTTP, no stdio conflict)",
+      "_setup": "Connect 500+ apps via Rube: Slack, Notion, Drive, GitHub, Granola, Otter.ai, Jira, Linear, Airtable, HubSpot, Stripe, and more. 1) Sign up free at https://rube.app 2) Connect your apps in Rube's marketplace 3) Copy your API key from the dashboard 4) Paste it in x-composio-api-key above 5) Restart Claude Code. See Rube section in CLAUDE.md for full guide."
     }
   },
 
   "_notes": {
     "memory": "Claudia's memory is powered by the claudia-memory daemon (Python MCP server). It provides ~33 tools for semantic search, pattern detection, and relationship tracking. The installer (npx get-claudia) automatically sets up the daemon in a Python venv at ~/.claudia/daemon/venv/ and configures .mcp.json.",
-    "gmail_and_calendar": "Gmail and Calendar use third-party MCP servers. Each user sets up their own Google Cloud Console credentials. See the Google Integration Setup section in CLAUDE.md.",
-    "rube": "Rube (by Composio) connects 500+ apps through one MCP server. Each user creates their own free Rube account at rube.app. See the Rube section in CLAUDE.md for setup and troubleshooting.",
+    "gmail_and_calendar": "Gmail and Calendar are enabled by default as stdio MCP servers. Each requires Google Cloud credentials (see Google Integration Setup in CLAUDE.md). Alternatively, connect them through Rube (HTTP) for a simpler setup.",
+    "rube": "Rube (by Composio) connects 500+ apps through one HTTP MCP connection. Each user creates their own free Rube account at rube.app. See the Rube section in CLAUDE.md for setup and troubleshooting.",
     "security": "Each user authenticates with their own accounts and credentials. OAuth tokens are stored locally on your machine, never shared.",
     "not_included": {
       "filesystem": "Claude Code has native Read/Write/Edit tools",
       "web-search": "Claude Code has native WebSearch",
       "fetch": "Claude Code has native WebFetch"
     },
-    "adding_servers": "Add MCP servers here if needed. Find more at mcp.so or github.com/modelcontextprotocol/servers"
+    "adding_servers": "Add MCP servers here if needed. Find more at mcp.so or github.com/modelcontextprotocol/servers."
   }
 }